agentengine-sdk-python 0.7.0__py3-none-any.whl → 0.8.0__py3-none-any.whl

ksadk/tracing/setup.py

@@ -5,12 +5,17 @@ Tracing 初始化 - 支持多 Exporter (InMemory + Langfuse + OTLP)
 
 import atexit
 import base64
+import importlib
 import logging
 import os
+from collections.abc import Callable, Sequence
 from dataclasses import dataclass
 from typing import Any, Optional
 from urllib.parse import unquote
 
+from opentelemetry.sdk.trace import ReadableSpan
+from opentelemetry.sdk.trace.export import SpanExporter, SpanExportResult
+
 from ksadk.tracing.exporters.inmemory_exporter import InMemoryExporter
 
 logger = logging.getLogger(__name__)
@@ -49,18 +54,18 @@ class _OtlpHttpConfig:
     service_name: str
 
 
-class _LoggingSpanExporter:
+class _LoggingSpanExporter(SpanExporter):
     """Small delegating exporter that logs external trace export flow."""
 
     def __init__(
         self,
-        exporter: Any,
+        exporter: SpanExporter,
         *,
         name: str,
         endpoint: str,
         service_name: str,
         header_keys: list[str],
-        span_transform: Any = None,
+        span_transform: Callable[[Sequence[ReadableSpan]], Sequence[ReadableSpan]] | None = None,
     ):
         self._exporter = exporter
         self._name = name
@@ -69,8 +74,8 @@ class _LoggingSpanExporter:
         self._header_keys = header_keys
         self._span_transform = span_transform
 
-    def export(self, spans):
-        span_count = len(spans) if hasattr(spans, "__len__") else "unknown"
+    def export(self, spans: Sequence[ReadableSpan]) -> SpanExportResult:
+        span_count = len(spans)
         logger.info(
             "%s export started: endpoint=%s service_name=%s spans=%s headers=%s",
             self._name,
@@ -101,24 +106,18 @@ class _LoggingSpanExporter:
         )
         return result
 
-    def shutdown(self):
-        shutdown = getattr(self._exporter, "shutdown", None)
-        if shutdown is None:
-            return None
+    def shutdown(self) -> None:
         logger.info("%s exporter shutdown: endpoint=%s", self._name, self._endpoint)
-        return shutdown()
+        self._exporter.shutdown()
 
     def force_flush(self, timeout_millis: int = 30000) -> bool:
-        force_flush = getattr(self._exporter, "force_flush", None)
-        if force_flush is None:
-            return True
         logger.info(
             "%s exporter force_flush: endpoint=%s timeout_millis=%s",
             self._name,
             self._endpoint,
             timeout_millis,
         )
-        return force_flush(timeout_millis)
+        return self._exporter.force_flush(timeout_millis)
 
 
 def _coerce_int(value: Any) -> int:
@@ -164,9 +163,7 @@ def _span_token_usage(span: Any) -> dict[str, int]:
         "input": _coerce_int(attributes.get("gen_ai.usage.input_tokens")),
         "output": _coerce_int(attributes.get("gen_ai.usage.output_tokens")),
         "cache_read": _coerce_int(attributes.get("gen_ai.usage.cache_read.input_tokens")),
-        "reasoning_output": _coerce_int(
-            attributes.get("gen_ai.usage.reasoning.output_tokens")
-        ),
+        "reasoning_output": _coerce_int(attributes.get("gen_ai.usage.reasoning.output_tokens")),
     }
 
 
@@ -294,13 +291,16 @@ def _prepare_cloud_monitor_spans(spans: Any) -> Any:
 
     def has_token_descendant(span: Any) -> bool:
         sid = _span_id(span)
-        key = (_trace_id(span), sid)
         if sid is None:
             return False
+        key = (_trace_id(span), sid)
         if key in has_token_descendant_cache:
             return has_token_descendant_cache[key]
         for child in children_by_parent.get(key, []):
-            child_key = (_trace_id(child), _span_id(child))
+            child_sid = _span_id(child)
+            if child_sid is None:
+                continue
+            child_key = (_trace_id(child), child_sid)
             if _has_nonzero_token_usage(token_usage_by_id.get(child_key, {})):
                 has_token_descendant_cache[key] = True
                 return True
@@ -312,6 +312,8 @@ def _prepare_cloud_monitor_spans(spans: Any) -> Any:
 
     def collect_leaf_usage(span: Any, usage: dict[str, int]) -> None:
         sid = _span_id(span)
+        if sid is None:
+            return
         key = (_trace_id(span), sid)
         own_usage = token_usage_by_id.get(key, {})
         if _has_nonzero_token_usage(own_usage) and not has_token_descendant(span):
@@ -349,7 +351,9 @@ def _prepare_cloud_monitor_spans(spans: Any) -> Any:
     return [replacements.get(id(span), span) for span in span_list]
 
 
-def _build_langfuse_otlp_config(langfuse_config: dict = None) -> Optional[dict]:
+def _build_langfuse_otlp_config(
+    langfuse_config: dict[str, Any] | None = None,
+) -> dict[str, Any] | None:
     """Build Langfuse OTLP direct exporter config from explicit config or env."""
     if langfuse_config:
         public_key = langfuse_config.get("public_key") or ""
@@ -358,7 +362,9 @@ def _build_langfuse_otlp_config(langfuse_config: dict = None) -> Optional[dict]:
     else:
         public_key = os.getenv("LANGFUSE_PUBLIC_KEY", "")
         secret_key = os.getenv("LANGFUSE_SECRET_KEY", "")
-        host = os.getenv("LANGFUSE_BASE_URL") or os.getenv("LANGFUSE_HOST") or "http://localhost:3000"
+        host = (
+            os.getenv("LANGFUSE_BASE_URL") or os.getenv("LANGFUSE_HOST") or "http://localhost:3000"
+        )
 
     if not public_key or not secret_key:
         return None
@@ -561,9 +567,7 @@ def _apply_langfuse_auth_fallback(endpoint: str, headers: dict[str, str]) -> boo
     langfuse_headers = _build_langfuse_otlp_headers(public_key, secret_key)
     headers["Authorization"] = langfuse_headers["Authorization"]
     if not _has_header(headers, "x-langfuse-ingestion-version"):
-        headers["x-langfuse-ingestion-version"] = langfuse_headers[
-            "x-langfuse-ingestion-version"
-        ]
+        headers["x-langfuse-ingestion-version"] = langfuse_headers["x-langfuse-ingestion-version"]
     return True
 
 
@@ -595,9 +599,8 @@ def _build_generic_otlp_http_config() -> Optional[dict]:
         )
         return None
 
-    raw_headers = (
-        os.getenv("OTEL_EXPORTER_OTLP_TRACES_HEADERS", "").strip()
-        or os.getenv("OTEL_EXPORTER_OTLP_HEADERS", "")
+    raw_headers = os.getenv("OTEL_EXPORTER_OTLP_TRACES_HEADERS", "").strip() or os.getenv(
+        "OTEL_EXPORTER_OTLP_HEADERS", ""
     )
     headers = _parse_otlp_headers(raw_headers)
     langfuse_auth_fallback = _apply_langfuse_auth_fallback(endpoint, headers)
@@ -618,13 +621,13 @@ def _build_generic_otlp_http_config() -> Optional[dict]:
 
 def setup_tracing(
     enable_inmemory: bool = True,
-    enable_langfuse: bool = None,  # Auto-detect from env
-    langfuse_config: dict = None,
+    enable_langfuse: bool | None = None,  # Auto-detect from env
+    langfuse_config: dict[str, Any] | None = None,
     enable_otlp: bool = False,
     otlp_endpoint: str = "localhost:4317",
     enable_adk_instrumentation: bool = True,  # Auto-instrument ADK
-    use_callback_only: bool = None,  # Explicit override
-    **kwargs
+    use_callback_only: bool | None = None,  # Explicit override
+    **kwargs,
 ) -> Optional[InMemoryExporter]:
     """初始化 Tracing (支持多 Exporter)
 
@@ -657,7 +660,7 @@ def setup_tracing(
 
     # 检查是否已有 TracerProvider (避免覆盖)
     existing_provider = trace.get_tracer_provider()
-    if existing_provider and hasattr(existing_provider, 'add_span_processor'):
+    if existing_provider and hasattr(existing_provider, "add_span_processor"):
         # 使用现有 provider，直接添加 processor
         provider = existing_provider
         logger.debug("Using existing TracerProvider")
@@ -679,14 +682,16 @@ def setup_tracing(
     generic_otlp_config = _build_generic_otlp_http_config()
     if generic_otlp_config:
         try:
-            from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
+            from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
+                OTLPSpanExporter as HttpOTLPSpanExporter,
+            )
 
-            otlp_exporter = OTLPSpanExporter(
+            raw_otlp_exporter = HttpOTLPSpanExporter(
                 endpoint=generic_otlp_config["endpoint"],
                 headers=generic_otlp_config["headers"],
             )
             otlp_exporter = _LoggingSpanExporter(
-                otlp_exporter,
+                raw_otlp_exporter,
                 name="Generic OTLP",
                 endpoint=generic_otlp_config["endpoint"],
                 service_name=_get_service_name(),
@@ -720,14 +725,16 @@ def setup_tracing(
     cloud_monitor_config = _build_cloud_monitor_otlp_http_config()
     if cloud_monitor_config:
         try:
-            from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
+            from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
+                OTLPSpanExporter as HttpOTLPSpanExporter,
+            )
 
-            cloud_monitor_exporter = OTLPSpanExporter(
+            raw_cloud_monitor_exporter = HttpOTLPSpanExporter(
                 endpoint=cloud_monitor_config.endpoint,
                 headers=cloud_monitor_config.headers,
             )
             cloud_monitor_exporter = _LoggingSpanExporter(
-                cloud_monitor_exporter,
+                raw_cloud_monitor_exporter,
                 name="CloudMonitor OTLP",
                 endpoint=cloud_monitor_config.endpoint,
                 service_name=cloud_monitor_config.service_name,
@@ -757,9 +764,8 @@ def setup_tracing(
     langfuse_enabled = enable_langfuse
     if langfuse_enabled is None:
         # Auto-detect from environment variables
-        langfuse_enabled = (
-            not generic_otlp_config
-            and bool(os.getenv("LANGFUSE_PUBLIC_KEY") or (langfuse_config or {}).get("public_key"))
+        langfuse_enabled = not generic_otlp_config and bool(
+            os.getenv("LANGFUSE_PUBLIC_KEY") or (langfuse_config or {}).get("public_key")
         )
 
     # 检查是否应该禁用 LangfuseExporter (当使用 LangChain/LangGraph 时)
@@ -769,16 +775,18 @@ def setup_tracing(
 
     if langfuse_enabled and not use_callback_only:
         try:
-            from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
+            from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
+                OTLPSpanExporter as HttpOTLPSpanExporter,
+            )
 
             config = _build_langfuse_otlp_config(langfuse_config)
             if config:
-                otlp_exporter = OTLPSpanExporter(
+                raw_langfuse_otlp_exporter = HttpOTLPSpanExporter(
                     endpoint=config["endpoint"],
                     headers=config["headers"],
                 )
                 otlp_exporter = _LoggingSpanExporter(
-                    otlp_exporter,
+                    raw_langfuse_otlp_exporter,
                     name="Langfuse OTLP",
                     endpoint=config["endpoint"],
                     service_name=_get_service_name(),
@@ -786,11 +794,11 @@ def setup_tracing(
                     span_transform=_prepare_langfuse_spans,
                 )
                 provider.add_span_processor(
-                BatchSpanProcessor(
-                    otlp_exporter,
-                    **_batch_span_processor_kwargs(),
+                    BatchSpanProcessor(
+                        otlp_exporter,
+                        **_batch_span_processor_kwargs(),
+                    )
                 )
-            )
                 logger.info(
                     "Langfuse OTLP exporter enabled: %s (%s) headers=%s",
                     config["endpoint"],
@@ -810,11 +818,14 @@ def setup_tracing(
     # 5. OTLP Exporter (optional)
     if enable_otlp:
         try:
-            from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import OTLPSpanExporter
-            otlp_exporter = OTLPSpanExporter(endpoint=otlp_endpoint)
+            from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import (
+                OTLPSpanExporter as GrpcOTLPSpanExporter,
+            )
+
+            grpc_otlp_exporter = GrpcOTLPSpanExporter(endpoint=otlp_endpoint)
             provider.add_span_processor(
                 BatchSpanProcessor(
-                    otlp_exporter,
+                    grpc_otlp_exporter,
                     **_batch_span_processor_kwargs(),
                 )
             )
@@ -825,7 +836,10 @@ def setup_tracing(
     # 6. ADK Auto-Instrumentation (for Google ADK projects)
     if enable_adk_instrumentation and not _adk_instrumented:
         try:
-            from openinference.instrumentation.google_adk import GoogleADKInstrumentor
+            instrumentation_module = importlib.import_module(
+                "openinference.instrumentation.google_adk"
+            )
+            GoogleADKInstrumentor = getattr(instrumentation_module, "GoogleADKInstrumentor")
             GoogleADKInstrumentor().instrument()
             _adk_instrumented = True
             logger.info("Google ADK instrumentation enabled")
@@ -841,6 +855,7 @@ def setup_tracing(
     if enable_adk_instrumentation:
         try:
             from openinference.instrumentation.langchain import LangChainInstrumentor
+
             LangChainInstrumentor().instrument()
             logger.info("LangChain instrumentation enabled")
         except ImportError:
@@ -864,7 +879,7 @@ def shutdown_tracing():
 
     if _langfuse_exporter is not None:
         try:
-            if hasattr(_langfuse_exporter, '_exporter'):
+            if hasattr(_langfuse_exporter, "_exporter"):
                 _langfuse_exporter._exporter.shutdown()
             logger.debug("Langfuse exporter shutdown gracefully")
         except Exception:
@@ -887,6 +902,7 @@ def get_tracer(name: str = "ksadk"):
     """获取 Tracer"""
     try:
         from opentelemetry import trace
+
         return trace.get_tracer(name)
     except ImportError:
         return None

ksadk/tracing/span_utils.py

@@ -93,9 +93,12 @@ def get_langfuse_callback(
         Langfuse callback handler 或 None
     """
     try:
-        from ksadk.tracing import get_langfuse_callback_handler
+        import ksadk.tracing as tracing_module
 
-        handler = get_langfuse_callback_handler()
+        factory = getattr(tracing_module, "get_langfuse_callback_handler", None)
+        if not callable(factory):
+            return None
+        handler = factory()
         if handler:
             # 设置 session 和 metadata
             handler.session_id = session_id

ksadk/tui/loop.py

@@ -3,6 +3,7 @@
 Transcript 使用 ANSI 保留 rich 颜色和 Markdown 落定格式，composer 紧跟内容，
 终端原生 scrollback 保留。支持键盘滚动、流式跟底和输入排队。
 """
+
 from __future__ import annotations
 
 import asyncio
@@ -105,7 +106,11 @@ def _codex_surface_rgb(
     # so 4% black on a light terminal is visually imperceptible; 8% gives the
     # input surface a stable edge without introducing a border.
     alpha = 0.08 if is_light and composer else 0.04 if is_light else 0.12
-    return tuple(int(top[i] * alpha + background[i] * (1.0 - alpha)) for i in range(3))
+    return (
+        int(top[0] * alpha + background[0] * (1.0 - alpha)),
+        int(top[1] * alpha + background[1] * (1.0 - alpha)),
+        int(top[2] * alpha + background[2] * (1.0 - alpha)),
+    )
 
 
 def _codex_surface_style(background: tuple[int, int, int] | None) -> str:
@@ -319,12 +324,16 @@ class SlashCommandCompleter(Completer):
             return
         # /model <prefix> → 补全可选模型 id
         if prefix.startswith("/model ") and self.runner is not None:
-            model_prefix = prefix[len("/model "):].strip()
+            model_prefix = prefix[len("/model ") :].strip()
             available = getattr(self.runner, "available_models", None) or []
             for m in available:
                 mid = str(m.get("id") or m.get("name") or "")
                 if mid and mid.startswith(model_prefix):
-                    yield Completion(mid, start_position=-len(model_prefix), display_meta=m.get("display_name") or "")
+                    yield Completion(
+                        mid,
+                        start_position=-len(model_prefix),
+                        display_meta=m.get("display_name") or "",
+                    )
             return
         for command, meta in self._COMMANDS.items():
             if command.startswith(prefix):
@@ -372,7 +381,7 @@ class RichLiveRenderer:
             self.assistant_entry.content = full_text
             self.assistant_entry.status = "streaming"
             if full_text.startswith(self._last_full_text):
-                delta = full_text[len(self._last_full_text):]
+                delta = full_text[len(self._last_full_text) :]
             else:
                 # A final/snapshot event can replace the cumulative text. Preserve
                 # ordering when possible; without tools it is safe to replace the
@@ -398,7 +407,9 @@ class RichLiveRenderer:
             self.assistant_entry.thinking = full_thinking
             # thinking 不进动态区，落定时随 assistant 一起渲染（show_thinking）
 
-    async def on_tool_call(self, tool_name: str, status: str, args: Any = None, call_id: str = "") -> None:
+    async def on_tool_call(
+        self, tool_name: str, status: str, args: Any = None, call_id: str = ""
+    ) -> None:
         if self.loop is None:
             return
         content = f"{tool_name} [{status}]"
@@ -441,7 +452,7 @@ class RichLiveRenderer:
     def final_entries(self, response: str) -> list[TranscriptEntry]:
         """Return finalized display cells in the original stream event order."""
         if response and response.startswith(self._last_full_text):
-            delta = response[len(self._last_full_text):]
+            delta = response[len(self._last_full_text) :]
             if delta:
                 if self._active_text_entry is None:
                     self._active_text_entry = TranscriptEntry(role="assistant")
@@ -507,7 +518,9 @@ class InteractionLoop:
         self._pin_to_bottom = True
         self._last_max_scroll = 0
         self._history_pager_active = False
-        self._entry_ansi_cache: dict[int, str] = {}  # 历史 entries ANSI 缓存（避免 streaming 全量重渲）
+        self._entry_ansi_cache: dict[int, str] = (
+            {}
+        )  # 历史 entries ANSI 缓存（避免 streaming 全量重渲）
         self._last_usage: dict[str, Any] | None = None
         self._showed_help = False  # 首次运行显示帮助列表
         self._welcome_ansi: str | None = None
@@ -600,6 +613,7 @@ class InteractionLoop:
             multiline=True,
             history=InMemoryHistory(),
         )
+        input_buffer = self._input_buffer
 
         # Codex uses an inline, content-first viewport: short transcripts keep
         # the composer directly below the content and a filler absorbs the rest
@@ -771,13 +785,13 @@ class InteractionLoop:
 
         @bindings.add("enter")
         def _submit(event) -> None:
-            text = self._input_buffer.text
-            self._input_buffer.reset(append_to_history=bool(text.strip()))
+            text = input_buffer.text
+            input_buffer.reset(append_to_history=bool(text.strip()))
             self._submit_text(text)
 
         @bindings.add("escape", "enter")
         def _newline(event) -> None:
-            self._input_buffer.insert_text("\n")
+            input_buffer.insert_text("\n")
 
         @bindings.add("c-c")
         def _interrupt_or_exit(event) -> None:
@@ -796,7 +810,7 @@ class InteractionLoop:
 
         @bindings.add("c-d")
         def _exit_on_empty(event) -> None:
-            if not self._input_buffer.text:
+            if not input_buffer.text:
                 event.app.exit()
 
         @bindings.add("pageup")
@@ -835,14 +849,18 @@ class InteractionLoop:
         @bindings.add("up", filter=self._picker_condition)
         def _picker_up(event) -> None:
             if self._model_picker_models:
-                self._model_picker_index = (self._model_picker_index - 1) % len(self._model_picker_models)
+                self._model_picker_index = (self._model_picker_index - 1) % len(
+                    self._model_picker_models
+                )
                 self._scroll_picker_to_selected()
                 event.app.invalidate()
 
         @bindings.add("down", filter=self._picker_condition)
         def _picker_down(event) -> None:
             if self._model_picker_models:
-                self._model_picker_index = (self._model_picker_index + 1) % len(self._model_picker_models)
+                self._model_picker_index = (self._model_picker_index + 1) % len(
+                    self._model_picker_models
+                )
                 self._scroll_picker_to_selected()
                 event.app.invalidate()
 
@@ -850,7 +868,11 @@ class InteractionLoop:
         def _picker_select(event) -> None:
             models = self._model_picker_models
             if models and 0 <= self._model_picker_index < len(models):
-                mid = str(models[self._model_picker_index].get("id") or models[self._model_picker_index].get("name") or "")
+                mid = str(
+                    models[self._model_picker_index].get("id")
+                    or models[self._model_picker_index].get("name")
+                    or ""
+                )
                 if mid:
                     self._apply_model_switch(mid)
                     self._close_model_picker()
@@ -980,7 +1002,9 @@ class InteractionLoop:
             for entry in renderer.final_entries(renderer._last_full_text):
                 await self._commit_entry(entry)
             if is_resume:
-                await self._commit_entry(TranscriptEntry(role="system", content="该 runtime 暂不支持审批续跑，已取消"))
+                await self._commit_entry(
+                    TranscriptEntry(role="system", content="该 runtime 暂不支持审批续跑，已取消")
+                )
                 self._clear_current_task()
                 self._drain_queue()
             else:
@@ -992,7 +1016,9 @@ class InteractionLoop:
             assistant_entry.status = ""
             for entry in renderer.final_entries(renderer._last_full_text):
                 await self._commit_entry(entry)
-            await self._commit_entry(TranscriptEntry(role="system", content="已取消（保留已产生内容）"))
+            await self._commit_entry(
+                TranscriptEntry(role="system", content="已取消（保留已产生内容）")
+            )
             self._clear_current_task()
             self._drain_queue()
             return
@@ -1018,7 +1044,9 @@ class InteractionLoop:
 
         async def _prompt():
             await self._commit_entry(
-                TranscriptEntry(role="system", content=f"确认执行 {tool_name}? 输入 y 确认，其他内容取消")
+                TranscriptEntry(
+                    role="system", content=f"确认执行 {tool_name}? 输入 y 确认，其他内容取消"
+                )
             )
 
         self._create_background_task(_prompt())
@@ -1032,7 +1060,9 @@ class InteractionLoop:
         if user_input.lower() in {"y", "yes"}:
 
             async def _resume():
-                await self._commit_entry(TranscriptEntry(role="system", content="已确认，尝试续跑..."))
+                await self._commit_entry(
+                    TranscriptEntry(role="system", content="已确认，尝试续跑...")
+                )
                 self._start_turn(
                     "",
                     input_data={**input_data, "resume": True},
@@ -1049,7 +1079,11 @@ class InteractionLoop:
             self._create_background_task(_cancel())
 
     def _drain_queue(self) -> None:
-        if self._pending_interrupt is not None or self._has_active_turn() or not self._queued_inputs:
+        if (
+            self._pending_interrupt is not None
+            or self._has_active_turn()
+            or not self._queued_inputs
+        ):
             return
         item = self._queued_inputs.pop(0)
         self._start_turn(item.text, user_entry=item.entry)
@@ -1167,7 +1201,7 @@ class InteractionLoop:
 
     def _handle_model_command(self, user_input: str) -> None:
         """/model:弹出交互式选择器(↑↓ Enter Esc)；/model <name>:直接切换(runnable 级,下轮生效)。"""
-        arg = user_input[len("/model"):].strip()
+        arg = user_input[len("/model") :].strip()
         if arg:
             # 直接切换：改 runner.model，下轮请求带新 model
             self._apply_model_switch(arg)
@@ -1192,7 +1226,9 @@ class InteractionLoop:
         """
         self.runner.model = model_id
         available = getattr(self.runner, "available_models", None) or []
-        matched = next((m for m in available if str(m.get("id") or m.get("name") or "") == model_id), None)
+        matched = next(
+            (m for m in available if str(m.get("id") or m.get("name") or "") == model_id), None
+        )
         if matched:
             self.runner.model_metadata = matched
         # 清旧观察名，让 _current_model_name 用新 runner.model
@@ -1263,8 +1299,8 @@ class InteractionLoop:
         frags: list[tuple[str, str]] = []
         for i, m in enumerate(self._model_picker_models):
             mid = str(m.get("id") or m.get("name") or "")
-            is_cur_model = (mid == current)
-            is_selected = (i == self._model_picker_index)
+            is_cur_model = mid == current
+            is_selected = i == self._model_picker_index
             marker = "›" if is_selected else " "
             line = f"{marker} {i + 1}. {mid}"
             if is_cur_model:
@@ -1276,7 +1312,7 @@ class InteractionLoop:
     # ---- streaming / transcript 渲染 ----
 
     def _set_streaming(self, full_text: str) -> None:
-        """streaming 文本更新到当前 streaming entry（动态区 = transcript 末尾的 streaming entry）。"""
+        """更新 transcript 末尾的 streaming entry。"""
         if self._streaming_entry is not None:
             self._streaming_entry.content = full_text
             self._streaming_entry.status = "streaming"
@@ -1365,6 +1401,7 @@ class InteractionLoop:
         height = int(getattr(ri, "window_height", 0) or 0)
         if height <= 0:
             import shutil
+
             height = max(10, (shutil.get_terminal_size(fallback=(80, 24)).lines or 24) - 6)
         line_count = self._transcript_line_count()
         return max(0, line_count - height)
@@ -1522,18 +1559,22 @@ def _resolve_model_name(runner) -> str:
         return str(meta["id"])
     observed = getattr(runner, "_observed_model", None)
     if observed:
-        return observed
-    return getattr(runner, "model", None) or os.getenv("MODEL_NAME") or "unknown"
+        return str(observed)
+    return str(getattr(runner, "model", None) or os.getenv("MODEL_NAME") or "unknown")
 
 
 def _help_text() -> str:
-    return "命令: /new 新会话 · /clear 清屏 · /session 查看 · /tools 展开工具 · exit 退出 · Ctrl-C 中断/退出"
+    return (
+        "命令: /new 新会话 · /clear 清屏 · /session 查看 · /tools 展开工具 "
+        "· exit 退出 · Ctrl-C 中断/退出"
+    )
 
 
 def _welcome_block(session_id: str, model_name: str, project_dir: Path, *, show_help: bool) -> str:
     """Render the compact Codex-style session card and one-line tip."""
     try:
         from ksadk.version import VERSION
+
         version = VERSION
     except Exception:
         version = ""
@@ -1552,6 +1593,7 @@ def _welcome_block(session_id: str, model_name: str, project_dir: Path, *, show_
     # 按终端宽度算框宽，留 transcript 左缩进 2 + 右 margin 2，上限 56（Codex 同款）。
     try:
         import shutil
+
         cols = shutil.get_terminal_size(fallback=(80, 24)).columns
     except Exception:
         cols = 80
@@ -1698,7 +1740,9 @@ def _render_entry_ansi(
             else:
                 console.print(Markdown(_normalize_markdown(f"{body}{suffix}")))
         else:
-            console.print(Text(suffix.lstrip(), style="dim") if suffix else Text("...", style="dim"))
+            console.print(
+                Text(suffix.lstrip(), style="dim") if suffix else Text("...", style="dim")
+            )
         return _strip_ansi_backgrounds(console.export_text(styles=True)).rstrip() + "\n"
 
     if entry.role == "tool":
@@ -1718,6 +1762,7 @@ def _render_entry_ansi(
             display_output = output
             try:
                 import json as _json
+
                 parsed = _json.loads(output)
                 # Tool adapters often JSON-encode an already serialized result.
                 # Unwrap that second layer so structured output folds by fields
@@ -1729,7 +1774,7 @@ def _render_entry_ansi(
                     if nested.startswith("content="):
                         import ast as _ast
 
-                        quote = nested[len("content="):len("content=") + 1]
+                        quote = nested[len("content=") : len("content=") + 1]
                         if quote in {"'", '"'}:
                             body_start = len("content=") + 1
                             body_end = body_start
@@ -1756,6 +1801,7 @@ def _render_entry_ansi(
                 pass
             # 估算显示行数：每行按 console width 折算（长字符串占多行）
             import shutil as _su
+
             cw = max(40, (_su.get_terminal_size(fallback=(80, 24)).columns or 80) - 6)
             logical_lines = display_output.split("\n")
             # 估算显示行数：按显示宽度（CJK 占 2）折算，非 len
@@ -1864,7 +1910,9 @@ def _strip_ansi_backgrounds(text: str) -> str:
     return re.sub(r"\x1b\[([0-9;]*)m", _clean, text)
 
 
-def run_tui(runner, *, show_thinking: bool = False, project_dir: str = ".", no_alt_screen: bool = False) -> None:
+def run_tui(
+    runner, *, show_thinking: bool = False, project_dir: str = ".", no_alt_screen: bool = False
+) -> None:
     """TUI 入口：被 cmd_invoke._invoke_tui / cmd_run._run_custom 调用。
 
     ``no_alt_screen`` 作为兼容参数保留；TUI 现在默认使用 Codex-style inline viewport。

ksadk/tui/stream_render.py

@@ -3,6 +3,7 @@
 从 RemoteRunner.stream 的归一化 chunk 提取 delta/usage/终止信号，格式化 usage 文本。
 被 loop.py（交互 TUI）和 cmd_invoke._invoke_once（-m 单次）共用，保证两路径渲染口径一致。
 """
+
 from __future__ import annotations
 
 import re